math64, timers: Fix 32bit mul_u64_u32_shr() and friends
It turns out that while GCC-4.4 manages to generate 32x32->64 mult
instructions for the 32bit mul_u64_u32_shr() code, any GCC after that
fails horribly.
Fix this by providing an explicit mul_u32_u32() function which can be
architcture provided.
Reported-by: Chris Metcalf <cmetcalf@mellanox.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Chris Metcalf <cmetcalf@mellanox.com> [for tile]
Cc: Christopher S. Hall <christopher.s.hall@intel.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Laurent Vivier <lvivier@redhat.com>
Cc: Liav Rehana <liavr@mellanox.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Parit Bhargava <prarit@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20161209083011.GD15765@worktop.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
diff --git a/arch/x86/include/asm/div64.h b/arch/x86/include/asm/div64.h
index ced283a..af95c47 100644
--- a/arch/x86/include/asm/div64.h
+++ b/arch/x86/include/asm/div64.h
@@ -59,6 +59,17 @@ static inline u64 div_u64_rem(u64 dividend, u32 divisor, u32 *remainder)
}
#define div_u64_rem div_u64_rem
+static inline u64 mul_u32_u32(u32 a, u32 b)
+{
+ u32 high, low;
+
+ asm ("mull %[b]" : "=a" (low), "=d" (high)
+ : [a] "a" (a), [b] "rm" (b) );
+
+ return low | ((u64)high) << 32;
+}
+#define mul_u32_u32 mul_u32_u32
+
#else
# include <asm-generic/div64.h>
#endif /* CONFIG_X86_32 */